Closing in on the gaps

Abstract

The case report in this issue of the Annals features a patient who developed metabolic acidosis following intravenous administration of propylene glycol contained in a standard preparation of lorazepam infusion. There is no specific symptom or sign of propylene glycol intoxication, and often the first clue lies in the finding of an unexpectedly high anion gap metabolic acidosis. The metabolic acidosis is attributed to the accumulation of lactic acid formed as a result of propylene glycol metabolism. The finding of a relatively normal serum lactic acid concentration in this case is surprising given also the underlying severe pneumonia. The reason for this, we speculate, was saturation of hepatic alcohol and aldehyde dehydrogenases preventing further metabolism to lactic acid, and removal of any lactic acid by the haemodialysis re-instated early in the course of her acute-on-chronic renal failure. Thus, it follows that the raised anion gap metabolic acidosis is caused independently by an increased endogenous production of anionic metabolites characteristic of hypercatabolic states, the renal failure per se and the introduction of anionic b-lactam given to the patient. Were it not for the serum osmolality measurement, leading to the discovery of a dramatic serum osmolal gap, the propylene glycol intoxication would have been overlooked. Thus, determination of the serum osmolal gap is recommended in cases of high anion gap metabolic acidosis of uncertain aetiology, particularly where intoxication by an alcohol such as ethanol, and in this case propylene glycol, cannot be excluded. We cannot stress enough the importance of a careful drug history in directing further toxicological investigations. The serum osmolal gap is nonetheless insensitive and lacks specificity in detecting an exogenous substance. A normal serum osmolal gap does not exclude the presence of a toxic substance since a small quantity of a low molecular weight substance such as ethanol may not raise the serum osmolal gap significantly. Furthermore, an initially high serum osmolal gap will gradually fall with time as the osmotically active parent compound is metabolised to polar metabolites which will in turn increase the serum anion gap. An increased serum osmolal gap is also seen in cases of diabetic ketoacidosis, circulatory shock, alcoholic acidosis and chronic renal failure. Analytical causes of a raised serum osmolal gap include sample contamination by EDTA or oxalate, and pseudohyponatraemia secondary to hyperproteinaemia or lipaemia. The latter is observed with indirect ion-selective electrodes which underestimate the true serum sodium concentration in such cases, leading to an apparently widened gap between the calculated serum osmolality and measured osmolality. Critics cited a lack of studies adequately validating its use in intoxicated patients, and the concept of subtracting a molarity concentration (mmol/L) from a measured osmolality (mOsm/kgH2O) is fundamentally flawed from a physicochemical perspective. Despite all these shortcomings, the serum osmolal gap is popular among clinicians because of its ready availability in the rapid detection of osmotically active substances in serum. Gas chromatography remains the method of choice in confirming alcohol intoxication, but its use is often restricted to tertiary centres with a less than 24 h service coverage. It is unfortunate that despite full clearance of the propylene glycol with extended haemodialysis, the patient deteriorated and died of ventilator-associated pneumonia shortly after. A developing high anion gap metabolic acidosis was again noted with similarly normal serum lactic acid concentration. We agree with the authors that this is likely due to accumulation of catabolic products associated with critical illness. Various workers have tried to improve upon the anion gap by measuring and incorporating more anions into the serum anion gap equation, narrowing the gap as it were. Albumin and phosphate, two of the next major ‘unmeasured’ anions omitted in the classical anion gap equation, have been incorporated into new equations with some application in critical care patients. These patients often have deranged serum albumin and phosphate concentrations which would have invalidated the classical anion gap equation. The notion of adding in a growing list of endogenous anions in the anion gap calculation should in theory improve the ability of the anion gap to detect any exogenous anions where a gap persists. Foreseeably the resultant equation can be rendered too complex and cumbersome for routine clinical use. This requires the need for a fundamental understanding of the basic pathophysiology of acidosis and the relevant factors that determine its generation. Stewart’s physicochemical approach did just that by identifying three independent parameters that determine the acid–base balance in any clinical scenario. Termed the strong ion difference concept, the three parameters are the difference between the strong cationic and anionic charges, the total concentrations of all the non-volatile